So, we are now in an era where “Big Data” matters. We have oodles of choice when it comes to storing this data and a myriad of languages and tools to help us extract and report on that data.
But what about monitoring. Do you know if you’ll run out of disk storage in the next month, next six months ? Do you know if your database is performing as well as it should ? Do you need more RAM to satisfy your query requirements ?
These are questions invariably asked when there is a problem. When a database is initially commissioned and it is all new and shiny, the focus is usually to get up and running. Who cares what might happen in six months’ time, just tell me what my customers are doing!!
And then disaster strikes. Six months down the road, space is getting tight, the data feeds are growing and you’ve committed to keeping 3 years worth of data on the database at any one time.
It’s a perfect storm…
I can think of numerous times in the last 20 years when the above scenario and has happened and, dare I say it, could have been avoided if monitoring had been in place. And it doesn’t have to be confined to databases. Any resource you rely on, can and will eventually run out of steam.
For KAP, we monitor various aspects of a database to ensure its running smoothly. When it isn’t, we are told by the monitoring system so that we can proactively do something about it.
So, if there is a predictive disk failure, we can, 99% of the time, fix it before it becomes an issue by hot-swapping out the failed disks. No downtime. No bother. Happy client.
Typically, we also monitor query queues, disk states, KAP backups, JDBC-ODBC bridge, slab management, hardware health for NICs, Port monitoring and VPN monitoring among others.
We can also make our plugins available to clients. These plugins are compatible with Nagios Core (for more information, click here : Nagios Core) which is widely used and open source.
There are many other monitoring systems out there that will also allow our plugins to be easily adapted if required.
We can also produce reports using addons from the monitoring data which gives us an insight on the usage of any given database. For example, disk storage can often be underestimated – whether it be because the data grew faster than expected or that the retention of the data was longer than expected – both can have an impact. By, monitoring disk usage, we can see just how quickly the disk storage is being used and look to try and increase that storage with minimal impact to the client. We can do the same for RAM.
We use pnp4nagios (for more information, click here : pnp4nagios) which is an open source graphing tool that utilizes the data generated by Nagios Core.